The positive results recently reported at the 1994 NAHS and ASTRO meetings from the European Phase III trials illustrate the continually increasing promise of hyperthermia as an adjuvant to radiation. One of the most promising hyperthermia heating methods is scanned focused ultrasound (SFU). Through obtaining good clinical temperatures while giving 595 treatments to 177 patients, we have demonstrated that an effective, practical (from the patient, clinical, engineering and economic standpoints) SFU hyperthermia system (SFUHS) can be designed, constructed and clinically implemented. This system has the potential for great flexibility in delivering power to tumors (e.g.: number, sizes, and frequencies of transducers; arbitrarily shaped 3D scanning paths; control of power along those paths); indeed, too much flexibility to be used optimally without both a significant clinical knowledge base and extensive computer assistance. Thus, due to both our initial time and clinical knowledge constraints, and to maintain consistency in our treatments, we have significantly limited several of the machine parameters in those first clinical studies, and only used select features of the SFUHS's flexibility. Nonetheless, where valid comparisons are possible our original system produced better temperature distributions than any other heating modality. Based on those clinical results we have identified several significant improvements that are needed to optimally use the full potential flexibility of our SFUHS. We propose to develop and add those improvements to our system and thereby remove one of the two most limiting constraints on the ability of all clinical systems to achieve good tumor temperature elevations-- the lack of flexibility in power deposition. The second such constraint is patient pain. We also propose to vigorously and methodically pursue methods of avoiding and alleviating patient pain, including both improved clinical procedures and a dose escalation study using general analgesia. The goals of this research are (l) to continue to improve our SFUHS and clinical procedures, and (2) to use this improved system to obtain higher, more uniform and controllable temperatures in a reduced overall treatment time. Specifically, we will develop and clinically test a series of continually improving optimal, learning, multi-input, multi-output feedback and feedforward control systems. We will systematically study the problem of patient pain by: identifying potential pain sites in pretreatment tests; optimizing power distributions through pretreatment planning; improving skin cooling; including pain and normal tissue temperatures in the controller; and administering general analgesics in a dose escalation study. We will perform all treatments using highly reproducible geometries, extensive thermometry, and known power depositions so that we can estimate both the intra- and inter-treatment relative blood flow and ultrasound absorption coefficient changes in patients. We will do this using site-specific, multi-treatment-window designs, including blood preheating. This will all be done in the context of a cooperative Phase II trial to determine if improved thermal doses result in improved clinical responses, and to test the alternate hypothesis that improved response with adjuvant hyperthermia is merely due to an unexplained association between tumor heatability and radiosensitivity.